Skip to main content

End-to-End, Transferable Deep RL for Graph Optimization

Privacy Considerations in Large Language Models

Portrait Light: Enhancing Portrait Lighting with Machine Learning

MediaPipe Holistic — Simultaneous Face, Hand and Pose Prediction, on Device

Google at NeurIPS 2020

Using AutoML for Time Series Forecasting

Transformers for Image Recognition at Scale

Navigating Recorder Transcripts Easily, with Smart Scrolling

The Language Interpretability Tool (LIT): Interactive Exploration and Analysis of NLP Models

Haptics with Input: Using Linear Resonant Actuators for Sensing

Using GANs to Create Fantastical Creatures

Mitigating Unfair Bias in ML Models with the MinDiff Framework

The Machine Learning Behind Hum to Search

Improving On-Device Speech Recognition with VoiceFilter-Lite

Announcing the Objectron Dataset

Background Features in Google Meet, Powered by Web ML

Experimenting with Automatic Video Creation from a Web Page

Estimating the Impact of Training Data with Reinforcement Learning

Rethinking Attention with Performers

Announcing the Recipients of the 2020 Award for Inclusion Research

Recreating Historical Streetscapes Using Deep Learning and Crowdsourcing

Measuring Gendered Correlations in Pre-trained NLP Models

Announcing the 2020 Google PhD Fellows

Massively Large-Scale Distributed Reinforcement Learning with Menger

Developing Real-Time, Automatic Sign Language Detection for Video Conferencing

Audiovisual Speech Enhancement in YouTube Stories

Advancing Instance-Level Recognition Research

Advancing NLP with Efficient Projection-Based Model Architectures

Improving the Accuracy of Genomic Analysis with DeepVariant 1.0

Improving Sparse Training with RigL

Imitation Learning in the Low-Data Regime

The Technology Behind our Recent Improvements in Flood Forecasting

KeyPose: Estimating the 3D Pose of Transparent Objects from Stereo

Using Machine Learning to Detect Deficient Coverage in Colonoscopy Screenings

Scaling Up Fundamental Quantum Chemistry Simulations on Quantum Hardware

Axial-DeepLab: Long-Range Modeling in All Layers for Panoptic Segmentation

An Analysis of Online Datasets Using Dataset Search (Published, in Part, as a Dataset)

Google at ECCV 2020

Understanding View Selection for Contrastive Learning

Tackling Open Challenges in Offline Reinforcement Learning

Understanding Deep Learning on Controlled Noisy Labels

Language-Agnostic BERT Sentence Embedding

On-device, Real-time Body Pose Tracking with MediaPipe BlazePose

REALM: Integrating Retrieval into Language Representation Models

A Simulation Suite for Tackling Applied Reinforcement Learning Challenges

On-device Supermarket Product Recognition

MediaPipe Iris: Real-time Iris Tracking & Depth Estimation

Live HDR+ and Dual Exposure Controls on Pixel 4 and 4a

Introducing the Model Card Toolkit for Easier Model Transparency Reporting

Announcing ScaNN: Efficient Vector Similarity Search

Improving Holistic Scene Understanding with Panoptic-DeepLab

Exploring Faster Screening with Fewer Tests via Bayesian Group Testing

Google at ICML 2020

Grounding Natural Language Instructions to Mobile UI Actions

AutoML-Zero: Evolving Code that Learns

Duality — A New Approach to Reinforcement Learning

Google at ACL 2020

SmartReply for YouTube Creators

SpineNet: A Novel Architecture for Object Detection Discovered with Neural Architecture Search

Leveraging Temporal Context for Object Detection

Sensing Force-Based Gestures on the Pixel 4

Presenting a Challenge and Workshop in Efficient Open-Domain Question Answering

RepNet: Counting Repetitions in Videos

Improving Speech Representations and Personalized Models Using Self-Supervision

Using Selective Attention in Reinforcement Learning Agents

Machine Learning-based Damage Assessment for Disaster Relief

Google at CVPR 2020

Extracting Structured Data from Templatic Documents

Unlocking the "Chemome" with DNA-Encoded Chemistry and Machine Learning

PEGASUS: A State-of-the-Art Model for Abstractive Text Summarization

Recent Advances in Google Translate

DADS: Unsupervised Reinforcement Learning for Skill Discovery

Federated Analytics: Collaborative Data Science without Data Collection

Evaluating Natural Language Generation with BLEURT

Open-Sourcing BiT: Exploring Large-Scale Pre-training for Computer Vision

Announcing the 7th Fine-Grained Visual Categorization Workshop

Enabling E-Textile Microinteractions: Gestures and Light through Helical Structures

Announcing Meta-Dataset: A Dataset of Datasets for Few-Shot Learning

Speeding Up Neural Network Training with Data Echoing

Agile and Intelligent Locomotion via Deep Reinforcement Learning

Understanding the Shape of Large-Scale Data

An NLU-Powered Tool to Explore COVID-19 Scientific Literature

Using Neural Networks to Find Answers in Tables

Applying Machine Learning to…..Yeast?

Yet More Google Compute Cluster Trace Data

Optimizing Multiple Loss Functions with Loss-Conditional Training

Google at ICLR 2020

Chip Design with Deep Reinforcement Learning

A Scalable Approach to Reducing Gender Bias in Google Translate

Exploring Evolutionary Meta-Learning in Robotics

Off-Policy Estimation for Infinite-Horizon Reinforcement Learning

EfficientDet: Towards Scalable and Efficient Object Detection

An Optimistic Perspective on Offline Reinforcement Learning

XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization

uDepth: Real-time 3D Depth Sensing on the Pixel 4

Advancing Self-Supervised and Semi-Supervised Learning with SimCLR

Exploring Nature-Inspired Robot Agility

Announcing the 2020 Image Matching Benchmark and Challenge

A Step Towards Protecting Patients from Medication Errors

Improving Audio Quality in Duo with WaveNetEQ

Exploring New Ways to Support Faculty Research

A Neural Weather Model for Eight-Hour Precipitation Forecasting

Massively Scaling Reinforcement Learning with SEED RL

Visual Transfer Learning for Robotic Manipulation

Introducing Dreamer: Scalable Reinforcement Learning Using World Models

Fast and Easy Infinitely Wide Networks with Neural Tangents

Soli Radar-Based Perception and Interaction in Pixel 4

Real-Time 3D Object Detection on Mobile Devices with MediaPipe

More Efficient NLP Model Pre-training with ELECTRA

Announcing TensorFlow Quantum: An Open Source Library for Quantum Machine Learning

Measuring Compositional Generalization

Toward Human-Centered Design for ML Frameworks

Ultra-High Resolution Image Analysis with Mesh-TensorFlow

Open Images V6 — Now Featuring Localized Narratives

Enhancing the Research Community’s Access to Street View Panoramas for Language Grounding Tasks

Exploring Transfer Learning with T5: the Text-To-Text Transfer Transformer

Announcing the 2019 Google Faculty Research Award Recipients

Setting Fairness Goals with the TensorFlow Constrained Optimization Library

Generating Diverse Synthetic Medical Image Data for Training Machine Learning Models

AutoFlip: An Open Source Framework for Intelligent Video Reframing

Learning to See Transparent Objects

TyDi QA: A Multilingual Question Answering Benchmark

ML-fairness-gym: A Tool for Exploring Long-Term Impacts of Machine Learning Systems

Encode, Tag and Realize: A Controllable and Efficient Approach for Text Generation

Announcing the Third Workshop and Challenge on Learned Image Compression

Towards a Conversational Agent that Can Chat About…Anything

Releasing the Drosophila Hemibrain Connectome — The Largest Synapse-Resolution Map of Brain Connectivity

Reformer: The Efficient Transformer

Can You Trust Your Model’s Uncertainty?

Using Machine Learning to “Nowcast” Precipitation in High Resolution

Google Research: Looking Back at 2019, and Forward to 2020 and Beyond